Novel Input and Output opportunitiesusing an Implanted Magnet

Paul StrohmeierSaarland University

Saarland Informatics Campusstrohmeier@cs.uni-saarland.de

Jess McIntoshUniversity of CopenhagenCopenhagen, Denmark

jemc@di.ku.dk

ABSTRACTIn this case study, we discuss how an implanted magnet can sup-port novel forms of input and output. By measuring the relativeposition between the magnet and an on-body device, local posi-tion of the device can be used for input. Electromagnetic fields canactuate the magnet to provide output by means of in-vivo hapticfeedback. Traditional tracking options would struggle tracking theinput methods we suggest, and the in-vivo sensations of vibrationprovided as output differ from the experience of vibrations appliedexternally – our data suggests that in-vivo vibrations are mediatedby different receptors than external vibration. As the magnet canbe easily tracked as well as actuated it provides opportunities forencoding information as material experiences.

CCS CONCEPTS• Human-centered computing → Human computer interac-tion (HCI).

KEYWORDSimplant, magnet, haptics, perception, implanted magnet

ACM Reference Format:Paul Strohmeier and Jess McIntosh. 2020. Novel Input and Output oppor-tunities using an Implanted Magnet. In AHs ’20: Augmented Humans Inter-national Conference (AHs ’20), March 16–17, 2020, Kaiserslautern, Germany.ACM, New York, NY, USA, 5 pages. https://doi.org/10.1145/3384657.3384785

1 INTRODUCTIONA magnetic implant can be used to enable input to and output inways which otherwise would be difficult or impossible to achieve.After briefly surveying related work, we discuss new input andoutput opportunities afforded by a magnetic implant1.

Implanted devices are not a new idea. Novels such as Gibson’sNeuromancer popularized the concept in the early 1980’s. Here,implants augment perception and abilities as well as providingusers with access to information. These fictional implants of the

1This paper is notmeant as an endorsement for implantingmagnets. The first author had themagnetimplanted in 2012, no in-vivo procedures were conducted for this paper.

Permission to make digital or hard copies of all or part of this work for personal orclassroom use is granted without fee provided that copies are not made or distributedfor profit or commercial advantage and that copies bear this notice and the full citationon the first page. Copyrights for components of this work owned by others than theauthor(s) must be honored. Abstracting with credit is permitted. To copy otherwise, orrepublish, to post on servers or to redistribute to lists, requires prior specific permissionand/or a fee. Request permissions from permissions@acm.org.AHs ’20, March 16–17, 2020, Kaiserslautern, Germany© 2020 Copyright held by the owner/author(s). Publication rights licensed to ACM.ACM ISBN 978-1-4503-7603-7/20/03. . . $15.00https://doi.org/10.1145/3384657.3384785

Figure 1: X-ray of the first author’s hand, palm facing thereader. The magnet is visible in white, located in the fattytissue above the bone.

past were interactive augmentations of our bodies, supersedingbodily abilities and at the forefront of our attention. This visiondid not become reality. Though implants permeate bodies in theform of stents, pacemakers, or medical delivery systems, they areat the background of our attention. Today’s implants are rarelyinteractive. Instead, the world around us has become more interac-tive, devices smart and responsive, and the use cases for interactiveimplants increasingly less obvious. The gritty future of hacked bod-ies is replaced by a present which provides interactivity withoutpenetrating the skin.

However, there are trends in HCI which suggest that we should –once again – consider the body more, exemplified by series of work-shops which explore merging bodies with technology [4, 13, 20]and discussion onHuman-Computer Integration [21]. Concurrently,DIY [2] and body-modification communities [29] have continuedexploring these ideas with functioning prototypes and basic im-planted devices such as magnets [27] or RFID chips [8]. It seemsprudent, then, to revisit the subject matter of implants.

We do so by (1) proposing implanted magnets used as local an-chors for on-body positioning, providing new opportunities forbody-centric interactions with wearable devices, and (2) highlight-ing that vibrotactile feedback provided by an implanted magnetis experienced differently from vibrotactile feedback applied ex-ternally, extending the potential bandwidth of haptic informationtransfer.

2 RELATEDWORKThere is a small body of work which has explored implanted inter-faces, however, devices which were actually implanted, typicallydo not take advantage of their unique placement with regards to

AHs ’20, March 16–17, 2020, Kaiserslautern, Germany P. Strohmeier and J. McIntosh

interaction. Instead, designers have opted for communicating withimplanted devices through mobile phones or other proxy devices[2]. Implanted devices could, however, be directly interacted with.Holz et al. [12] demonstrated that, theoretically, implanted devicescould provide a broad variety of input options, however their workdoes not consider biocompatability.

Implanted devices have also been used for measuring neuralactivity, for example, by inserting sensing-meshes between theskull and the brain, to collect high-definition real time informationof neural activities. Typically, such systems however are deployedto collect data which is later analyzed post-hoc, not affording anyreal time interaction [6]. An interesting twist to the ‘implanted’genre are devices which are swallowed, as demonstrated by Stelarc[26] and more recently by Li et al. [18]. However, the device byStelarc was intended to be viewed with medical imagery devices,while the device by Li et al. is interacted with through a smart-phone app. Again, neither device takes advantage of its in-vivoposition in terms of novel interactions.

While a range of in-vivo devices exist, the question of how onemight use such devices for novel interaction is still largely unan-swered. A simple implant, which can easily support novel interac-tions is a magnet (see also Heffernan et al. [11]). Such an implantedmagnet has no active parts which require power or control, andcan provide input to magnetic tracking systems, and can be providethe user with in-vivo haptic stimuli.

The potential for novel interactions using an implanted magnetshould be obvious, when looking at the plethora of HCI prototypeswhich currently improve their input abilities using a magnet [1, 9,19] and the common use of solenoid style actuators for providinghaptic feedback [16, 24, 28].

3 SYSTEM COMPONENTSThe interactions presented use a wearable device, augmented withMagnetips, and an implanted magnet:

Magnetips: The hardware used was previously published as Mag-netips – a system which could track and actuate a magnet remotely[19]. Magnetips was in part inspired by prototypes built to vibratethe magnetic implant of the first author. The interest in trackingthe location of the magnet originated in part from the wish to dolocal on-body position tracking of devices relative to the implant’slocation. Here we extend upon our previous work by discussinghow magnetips interacts with an implanted devices.

Implanted Magnet: The first author of this paper has a magnetimplanted in the palm of their left hand (see Figure 1). The magnetis a neodymium magnet approximately 2mm in diameter and 4mmlength. The magnet is coated with Parylene C. The position waschosen as it is relatively well protected by fatty tissue, and notlikely to significantly interfere with equipment which is sensitiveto magnetism. The palm is also an area of the hand with relativelystrong sensory innervation. While the fingertip would offer evenstronger innervation [14], positioning the magnet there has thepotential to interfere with activities such as climbing, or calibratingequipment sensitive to magnetism.

The interactions presented use a smartwatch as a familiar exam-ple. The Magnetips system, however, might instead be embeddedin garments [5] or jewelry [22], or might be used as guidance andinteraction system for epidermal robots roaming the body [3].

Distance between device and magnet

Magnet

Figure 2: The position of the smartwatch can be inferredby its distance to the implanted magnet. Here, positions aremapped to applications. Close to the wrist, the smartwatchdisplays the time, when moved away from the wrist the dis-play switches to a map. This can also be used for continu-ous input. For example, the user might hold the watch-facesteady while rotating their hand to scroll through a list.

4 INPUTMagnetips enables tracking of a magnet attached to the user’sfingernail. Instead of using the system to interact with a magnetattached to a fingernail [19], Magnetips can also be used to interactwith an implanted magnet. Depending on how the Magnetips sys-tem is attached to the body, different types of interaction becomepossible:

4.1 (a) Device ManipulationsIf the wearable device is loosely fit around the body, the magnet canbe used as an anchor-point for detecting local on-body position ofthe wearable device. This can enable a user to change the device’sposition or orientation as an input-method. For example sliding thedevice up or down the wrist (see Figure 2), or rotating it aroundthe wrist.

Rotation based interactions might be captured using an InertialMeasurement Units (IMU), however, an IMU cannot distinguishbetween a rotation of the users arm together with the device, androtating the device while the arm remains stationary. Additionally,translation (for example moving a bracelet towards or away fromthe wrist) would be hard to capture with an IMU at all. In Figure2, a user selects applications based on the devices position on thearm. Position is inferred by the distance to the magnet.

22:45 23:45

Angle between device and magnet

Magnet

Figure 3: If the smartwatch is in a fixed position, any changein direction of the magnet can be attributed to hand move-ment. This enables robust gestural interaction.Here the userhas applications mapped to hand-poses. A user switchingfrom a time-telling tomap application, based onwrist-angle.Continuous movement might also be used for continuouscontrol, such as volume adjustment.

Novel Input and Output Opportunities using an Implanted Magnet AHs ’20, March 16–17, 2020, Kaiserslautern, Germany

80 Hz 160 Hz 320 Hz 640 Hz

In-V

ivo

vibr

atio

n

500 Hz

333 Hz

250 Hz167 Hz

83 Hz

80 Hz 160 Hz 320 Hz 640 Hz

Exte

rnal

vib

ratio

n

80 Hz 160 Hz 320 Hz 640 Hz

1 db (80 Hz)

-12 db (160 Hz) -17.5 db

(250 Hz) - 16 db (320 Hz)

- 8 db (640 Hz)

Det

ectio

n Th

resh

olds

a b c

Figure 4: (a) Vibration detection thresholds by Verillo [32], yellow line shows the intensity of vibration required for it to beperceived. (b) Magnitude estimation study by McIntosh et al. [19], red lines show estimates of perceived strength of vibration,blue line represents the mean. (c) Replication of the experiment conducted by McIntosh et al. but with single participant andan implanted magnet. Green dots show estimated magnitudes, green line shows polynomial function fit to the data.

4.2 (b) Body MovementsIf Magnetips is in a known position (for example by being firmlyconnected to a body-part) all measuredmovement of themagnet canbe attributed to movements of the body. For example, if Magnetipsis firmly attached to the wrist, and the magnet is implanted in theusers hand, all motion of the magnet must be caused by motionof the users hand. This can be used to support gestural input. InFigure 3 the user selects applications based on wrist angle. Wristangle is inferred based on the direction of the magnet.

4.3 (c) Mixed Body-Centric InteractionsMagnetips also has an IMU, so it can detect its own movements. Insum – given the position of the first authors implant – Magnetipscan distinguish between three scenarios. If the smartwatch moves,but the magnet remains in a fixed relative position, we can infermovement of the arm. If the smartwatch moves and the magnetperforms the opposite relative motion, we can infer movement ofthe device. Finally, if the smartwatch is stationary, but the magnetmoves, we can infer movement of the hand. This allows combiningarm movements, wrist movements and device manipulation, pro-viding a rich space for body-relative interactions, which otherwisewould be difficult to achieve.

5 OUTPUTMagnetips can also actuate a magnet remotely – be it an external orimplanted magnet. Previous work has already suggested that usersare more sensitive to in-vivo vibrations, and has speculated thatthe mediation methods between the two types of stimulation mightdiffer [10]. We extend upon this work by exploring the experienceof in-vivo vibration over a wider frequency range.

Sections 5.1 and 5.2 present two experiments which were con-ducted by the second author with the first author as participant.The first author had no information regarding order of conditions,and was not aware of the parameters of the stimuli he felt. It shouldbe pointed out that within psycho-physics, such single-participantexperiments are not uncommon for establishing relations betweenphysical stimulus and experienced sensation [7]. Single participantcase-studies are also common within HCI (see also Lazar et al.,Chapter 7 [17]). For better comparison of results, we use the samehardware and software which was used in the evaluation of Mag-netips [19]. The experiments were conducted two years ago outof pure curiosity, at the time we did not expect these results andhad no intention of writing this paper, so there was no particulardesired outcome which might have biased the results.

5.1 Surprising ObservationsIn this section we compare results from three experiments: (a)measures of human sensitivity to vibration by Verillo [32] (Figure4a), (b) our prior magnitude estimation study using Magnetips [19](Figure 4b – overall average in blue, 10 individual participants inred), and (c) replication of the experiment conducted by McIntoshet al. but with an implanted magnet (Figure 4c). A summary of thesestudies can be found in Table 1.

Verillo found that the frequency which required the minimumamplitude to be perceived was at ∼250Hz. For higher or lowerfrequencies, participants required a higher amplitude to perceivethe vibrations (Figure 4a) [32]. In the Magnetips study we askedparticipants to report the perceived magnitude of a stimulus atvarying frequency (Figure 4b). We expected to find a result nega-tively correlated to the data provided by Verillo, which we did. Thereported magnitudes in the Magnetips study largely correspondsto the detection thresholds observed by Verillo. There appears tobe a discrepancy in the peak of the two signals (250 Hz and 33 Hzrespectively). This could be due to measurement error or idiosyn-crasies of the Magnetips system. It should be noted that a resultwhich correlates perfectly with the data by Verillo [32] would notbe significantly different from our data at p < 0.05 (see [19]).

Verillo [33] later showed that his measures correspond closelyto direct cell readings of Pacinian Corpuscles obtained by Sato [25].It can be assumed that the vibrations created using Magnetips arealso mediated by Pacinian Corpuscles, though an explanation forthe discrepancy in peak sensitivity requires further investigation.

We repeated the magnitude estimation task presented in Mag-netips [19], with the first author, who has an implanted magnet, assole participant (Figure 4c). We found similar effects of duration asbefore, but the effects of frequency were markedly different. Weplot raw response scores as green circles and – as we would expectto find a polynomial function – we superimposed the best fittingpolynomial function of second degree on our results (Figure 4c).In the high frequency areas – above ∼300Hz– the results are asexpected, corresponding both to the experiment by McIntosh (Fig-ure 4b [19]) and with what we would expect based on detection

Method Vibration Participant #

(a) Verillo [32] DetectionThreshold External 3

(b) McIntosh [19] MagnitudeEstimation External 10

(c) Present Studies MagnitudeEstimation In Vivo Single

Table 1: Overview of studies (studies b and c used the sameexperimental hardware and software).

AHs ’20, March 16–17, 2020, Kaiserslautern, Germany P. Strohmeier and J. McIntosh

16 Hz

32 Hz

56 Hz

88 Hz

128 Hz

176 Hz

232 Hz

296 Hz

368 Hz

448 Hz

10 Hz 100 Hz 1000 Hz

16 Hz

32 Hz

56 Hz

88 Hz

128 Hz

176 Hz

232 Hz

296 Hz

368 Hz

448 Hz

10 Hz 100 Hz 1000 Hz

16 Hz

32 Hz

56 Hz

88 Hz

128 Hz

176 Hz

232 Hz

296 Hz

368 Hz

448 Hz

10 Hz 100 Hz 1000 Hz

a b cPerc

eive

d St

reng

th

Perc

eive

d Sh

arpn

ess

Perc

eive

d D

efor

mat

ion

Figure 5: Magnitude estimation of in-vivo vibration: (a) estimates of perceived strength, (b) estimates of perceived sharpnessand (c) estimates of perceived deformation. Each point represents one estimate, solid lines are the polynomial function whichbest fits the recorded data. Expected sensitivity to vibration added yellow for reference (the inverse of detection thresholdsreported by Verillo [32]).

thresholds (Figure 4a [32]). However, for low frequencies – below250Hz– the perceived intensity increased, which is surprising.

5.2 Experiencing In-Vivo VibrationAs the relatively high strength of low frequency vibration wasunexpected, we conducted further, informal tests with a wider pa-rameter range and different pulse durations. We found that varyingthe frequency created qualitatively distinct experiences. While de-scribing these beyond that they are distinct is difficult, some ofthese experiences felt sharp, while others felt less like vibrationand more as if something were pulling or pushing. To see if theunexpected effect of frequency was consistent, and to gain insighton the qualitative experience, we repeated the experiment. Thistime we tested a wider range of frequencies (16Hz, 32Hz, 56Hz,88Hz, 128Hz, 176Hz, 232Hz, 296Hz, 368Hz, 448Hz) but only asingle level of pulse duration (1200ms). We chose the frequencylevels so that the differences between the stimuli were a geometricseries, and so they covered the range we were most interested in.Stimulus order was randomized and each frequency was presentedfour times. The entire procedure was repeated three times, oncefor perceived strength, once for perceived sharpness and once forperceived deformation (experience of pulling or pushing).

The results are plotted in Figure 5a. Raw estimates are indicatedas green circles and the best fitting polynomial function of seconddegree as a solid green curves. The expected response curve forperceived strength, based on the inverse of detection thresholdsestablished by Verillo [32], is indicated in yellow.

We found that, again, the perceived strength is strongest for lowfrequencies, below 250Hz (Figure 5a). This is surprising, as we ex-pected the experienced strength to drop off similarly below 250Hzas it does above 250Hz (indicated in yellow). One might argue thatthere is an observed peak at 88Hz and then a slight decline belowthat, but further experimentation is required for stating this withany certainty. This result is at odds with the Magnetips study (seealso Figure 4b) [19] and to what one would expect based on detec-tion threshold alone (Figure 4a) [32], but consistent with our initialexperiment using the implant (Figure 4c).

The experience of sharpness (Figure 5b) had a very slight posi-tive correlation with frequency, similar to previous studies whichrelated frequency to sharpness [30]. The experience of deformation(Figure 5c) was strongest at low frequencies (∼100Hz and lower)and decreased at higher frequencies. We assume that this experi-ence might be the reason that the overall magnitude estimationresults for the implanted magnet differ so strongly from what we

expected to find. There is a strong difference in quality of experi-ence below ∼100Hz (high deformation, low sharpness) comparedto above 100Hz (low deformation, high sharpness).

5.3 Discussion of Haptics ExplorationsThe difference between the results found in the in-vivo study andthe previous study by McIntosh et al. [19] might be attributed to thelocation of the stimulation. The study by McIntosh stimulated thefingernail, while here the stimulation occurred in the palm of thehand. The study by Verillo [32], however, also stimulated the palmof the hand, which makes this explanation unlikely. One might alsoargue that the deformation experience might be explained by themagnet literally being pushed and pulled. While this explanationis possible, we also find it unlikely as the stimulation mechanismwas the same as used by McIntosh et al. [19]. Also the experienceof deformation was not of a rapid pushing and pulling, but rather aslow, continuous sensation.

While the data is too preliminary to draw any strong conclu-sions beyond the observation that the results are unexpected, areasonable argument can be made for the in-vivo magnet stim-ulating receptors differently to externally applied vibration: Thesensitivity of Pacinian cells, which are usually associated with theperception of textures and vibration, peaks between ∼250Hz and∼300Hz and decreases linearly above and below that [14]. Thisaligns with the measures by Verillo [32, 33] and the results of thestudy by McIntosh et al. [19]. Meissner Corpuscles are responsive tolower frequency vibration from ∼5Hz to ∼50Hz (sometimes alsoreferred to as flutter-vibration [31]). They are similar in structureto Pacinian Corpuscles and typically associated with the perceptionof fine surface features, edges and contours [14]. The observedresults would make sense if in-vivo vibrations were more stronglymediated by Meissner corpuscles than externally applied vibrations.

6 TOWARDS FUTURE APPLICATIONSThe implanted magnet provides opportunities beyond body-relativeinput and corresponding vibrotactile feedback. Co-located sensingand actuation is promising for conveying a variety of material expe-riences: recent explorations of haptic experiences such as textures[24], compliance [15], abstract mid-air sensations [28], or concretemediated forces, such as penetrating a drop of water with a needle[23], not only rely on precise haptic actuation, but in same partsrequire precise tracking of human actions. If we wish to extend thehumans perceptive horizon by providing information we usually

Novel Input and Output Opportunities using an Implanted Magnet AHs ’20, March 16–17, 2020, Kaiserslautern, Germany

do not experience directly, such as the presence of a WiFi signalor the passing of time, we could use similar schemes for encodingthis information. Such encoding would require both measurementof movements, and precise actuation which can be achieved usingthe implanted magnet.

In anecdotal reports of sensations conveyed through implantedmagnets, it is often highlighted that it takes weeks or even monthsbefore users can correctly interpret the sensations created whenthe magnet is stimulated by arbitrary electromagnetic fields in onesenvironment [11, 27]. This highlights that the permanent placementof the magnet might support the user in creating a mental modelfor interpreting the sensation.

7 CONCLUSIONWe have demonstrated that an implanted magnet can provide inputopportunities for HCI which are difficult or impossible to achieveusing more traditional means such as an IMU. In terms of outputprovided through the implanted magnet, the clear qualitative dif-ferences between external and in-vivo vibration might potentiallybe leveraged to provide a wider bandwidth of haptic informationto a user. The experienced strength of of low-frequency in-vivovibration invites further investigation.

REFERENCES[1] Daniel Ashbrook, Patrick Baudisch, and Sean White. 2011. Nenya: subtle and

eyes-free mobile input with a magnetically-tracked finger ring. In Proceedings ofthe SIGCHI Conference on Human Factors in Computing Systems. ACM, 2043–2046.

[2] Lauren M. Britton and Bryan Semaan. 2017. Manifesting the Cyborg ThroughTechno-Body Modification: From Human-Computer Interaction to Integration.In Proceedings of the 2017 CHI Conference on Human Factors in Computing Systems(CHI ’17). ACM, New York, NY, USA, 2499–2510. https://doi.org/10.1145/3025453.3025629

[3] Artem Dementyev, Javier Hernandez, Inrak Choi, Sean Follmer, and JosephParadiso. 2018. Epidermal Robots: Wearable Sensors That Climb on the Skin.Proc. ACM Interact. Mob. Wearable Ubiquitous Technol. 2, 3, Article 102 (Sept.2018), 22 pages. https://doi.org/10.1145/3264912

[4] Ellen Yi-Luen Do, Kristina Höök, Pattie Maes, and Florian Mueller. [n. d.]. De-signing the Human-Machine Symbiosis. https://www.dagstuhl.de/en/program/calendar/semhp/?semnr=20272

[5] Rachel Freire, Cedric Honnet, and Paul Strohmeier. 2017. Second Skin: AnExploration of eTextile Stretch Circuits on the Body. In Proceedings of the EleventhInternational Conference on Tangible, Embedded, and Embodied Interaction. ACM,653–658.

[6] Tian-Ming Fu, Guosong Hong, Robert D. Viveros, Tao Zhou, and Charles M.Lieber. 2017. Highly scalable multichannel mesh electronics for stablechronic brain electrophysiology. Proceedings of the National Academy of Sci-ences 114, 47 (2017), E10046–E10055. https://doi.org/10.1073/pnas.1717695114arXiv:https://www.pnas.org/content/114/47/E10046.full.pdf

[7] George A. Gescheider. 1997. Psychophysics: the fundamentals. L. Erlbaum Asso-ciates. 435 pages.

[8] Amal Graafstra. 2007. Hands On: How radio-frequency identifion and I gotpersonal. IEEE Spectrum March (2007), 18–23.

[9] Chris Harrison and Scott E. Hudson. 2009. Abracadabra: Wireless, High-precision,and Unpowered Finger Input for Very Small Mobile Devices. In Proceedings of the22Nd Annual ACM Symposium on User Interface Software and Technology (UIST’09). ACM, New York, NY, USA, 121–124. https://doi.org/10.1145/1622176.1622199

[10] Ian Harrison, Kevin Warwick, and Virginie Ruiz. 2018. Subdermal MagneticImplants: An Experimental Study. Cybernetics and Systems 49, 2 (2018), 122–150.https://doi.org/10.1080/01969722.2018.1448223

[11] Kayla J. Heffernan, Frank Vetere, and Shanton Chang. 2016. You Put What,Where?: Hobbyist Use of Insertable Devices. In Proceedings of the 2016 CHIConference on Human Factors in Computing Systems (CHI ’16). ACM, New York,NY, USA, 1798–1809. https://doi.org/10.1145/2858036.2858392

[12] Christian Holz, Tovi Grossman, George Fitzmaurice, and Anne Agur. 2012. Im-planted User Interfaces. In Proceedings of the SIGCHI Conference on HumanFactors in Computing Systems (CHI ’12). ACM, New York, NY, USA, 503–512.https://doi.org/10.1145/2207676.2207745

[13] Kasper Hornbaek, David Kirsh, Joseph A. Paradiso, and Jürgen Steimle. 2018.On-Body Interaction: Embodied Cognition Meets Sensor/Actuator Engineering

to Design New Interfaces (Dagstuhl Seminar 18212). Dagstuhl Reports 8, 5 (2018),80–101. https://doi.org/10.4230/DagRep.8.5.80

[14] Roland S Johansson and J Randall Flanagan. 2009. Coding and use of tactile signalsfrom the fingertips in object manipulation tasks. Nature reviews. Neuroscience 10,5 (2009), 345–59. https://doi.org/10.1038/nrn2621

[15] Johan Kildal. 2010. 3D-press: Haptic Illusion of Compliance when Pressing ona Rigid Surface. In International Conference on Multimodal Interfaces and theWorkshop on Machine Learning for Multimodal Interaction (ICMI-MLMI ’10). ACM,New York, NY, USA, Article 21, 8 pages. https://doi.org/10.1145/1891903.1891931

[16] Hwan Kim, HyeonBeom Yi, Hyein Lee, and Woohun Lee. 2018. HapCube: AWearable Tactile Device to Provide Tangential and Normal Pseudo-Force Feed-back on a Fingertip. In Proceedings of the 2018 CHI Conference on Human Factorsin Computing Systems (CHI ’18). ACM, New York, NY, USA, Article 501, 13 pages.https://doi.org/10.1145/3173574.3174075

[17] Jonathan Lazar, JinjuanHeidi Feng, andHarryHochheiser. 2017. ResearchMethodsin Human-Computer Interaction. Elsevier Inc. 1–560 pages. https://doi.org/10.1016/b978-044481862-1/50075-3

[18] Zhuying Li, Yan Wang, Wei Wang, Weikang Chen, Ti Hoang, Stefan Greuter,and Florian Floyd Mueller. 2019. HeatCraft: Designing Playful Experiences withIngestible Sensors via Localized Thermal Stimuli. In Proceedings of the 2019 CHIConference on Human Factors in Computing Systems (CHI ’19). ACM, New York,NY, USA, Article 576, 12 pages. https://doi.org/10.1145/3290605.3300806

[19] Jess McIntosh, Paul Strohmeier, Jarrod Knibbe, Sebastian Boring, and KasperHornbæk. 2019. Magnetips: Combining Fingertip Tracking and Haptic Feedbackfor Around-Device Interaction. In Proceedings of the 2019 CHI Conference onHuman Factors in Computing Systems (CHI ’19). ACM, New York, NY, USA, Article408, 12 pages. https://doi.org/10.1145/3290605.3300638

[20] Florian Mueller, Pattie Maes, and Jonathan Grudin. 2019. Human-ComputerIntegration (Dagstuhl Seminar 18322). Dagstuhl Reports 8, 8 (2019), 18–47. https://doi.org/10.4230/DagRep.8.8.18

[21] Florian Floyd Mueller, Pedro Lopes, Paul Strohmeier, Wendy Ju, Caitlyn Seim,MartinWeigel, Suranga Nanayakkara, Marianna Obrist, Zhuying Li, Joseph Delfa,Jun Nishida, Elizabeth M Gerber, Dag Svanaes, Jonathan Grudin, Stefan Greuter,Kai Kunze, Thomas Erickson, Steven Greenspan, Masahiko Inami, Joe Marshall,Harald Reiterer, Katrin Wolf, Johen Meyer, Thecla Schiphorst, Dakuo Wang, andPattie Maes. 2020. Next Steps in Human-Computer Integration. In Proceedingsof the 2019 CHI Conference on Human Factors in Computing Systems (CHI ’19).https://doi.org/10.1145/3313831.3376242

[22] Karin Niemantsverdriet and Maarten Versteeg. 2016. Interactive JewelleryAs Memory Cue: Designing a Sound Locket for Individual Reminiscence. InProceedings of the TEI ’16: Tenth International Conference on Tangible, Embed-ded, and Embodied Interaction (TEI ’16). ACM, New York, NY, USA, 532–538.https://doi.org/10.1145/2839462.2856524

[23] Abdenbi Mohand Ousaid, Guillaume Millet, Sinan Haliyo, Stéphane Régnier, andVincent Hayward. 2014. Feeling what an insect feels. PLoS ONE 9, 10 (2014),e108895. https://doi.org/10.1371/journal.pone.0108895

[24] Joseph M. Romano and Katherine J. Kuchenbecker. 2012. Creating RealisticVirtual Textures from Contact Acceleration Data. EEE Trans. Haptics 5, 2 (Jan.2012), 109–119. https://doi.org/10.1109/TOH.2011.38

[25] M Sato. 1961. Response of Pacinian Corpuscles to Sinusoidal Vibrations. TechnicalReport. 391–409 pages. https://physoc.onlinelibrary.wiley.com/doi/pdf/10.1113/jphysiol.1961.sp006817

[26] Stelarc. 1993. Stomach Sculpture. , http://stelarc.org/?catID=20349 pages. https://aboutstelarc.weebly.com/stomach-sculpture.html

[27] Paul Strohmeier. 2013. Magnetic Implant & Sensing Electromagnetic Fields.http://fkeel.blogspot.com/2013/01/magnetic-implant-sensing.html

[28] Paul Strohmeier, Sebastian Boring, and Kasper Hornbæk. 2018. From Pulse Trainsto "Coloring with Vibrations": Motion Mappings for Mid-Air Haptic Textures. InProceedings of the 2018 CHI Conference on Human Factors in Computing Systems(CHI ’18). ACM, New York, NY, USA, Article 65, 13 pages. https://doi.org/10.1145/3173574.3173639

[29] Paul Strohmeier, Cedric Honnet, and Samppa Von Cyborg. 2016. Developing anEcosystem for Interactive Electronic Implants. Proc. Living Machines 2016 (2016).https://doi.org/10.1007/978-3-319-42417-0

[30] Paul Strohmeier and Kasper Hornbæk. 2017. Generating Haptic Textures witha Vibrotactile Actuator. In Proceedings of the 2017 CHI Conference on HumanFactors in Computing Systems (CHI ’17). ACM, New York, NY, USA, 4994–5005.https://doi.org/10.1145/3025453.3025812

[31] WHTalbot, I Darian-Smith, HHKornhuber, and VBMountcastle. 1967. The senseof flutter-vibration: comparison of the human capacity with response patterns ofmechanoreceptive afferents from the monkey hand. Journal of Neurophysiology31, 2 (1967), 301–334. https://doi.org/10.1152/jn.1968.31.2.301

[32] Ronald T Verrillo. 1962. Effect of Contactor Area on the Vibrotactile Threshold.The Journal of the Acoustical Society of America 35, 12 (1962). https://doi.org/10.1121/1.1918868

[33] Ronald T. Verrillo. 2014. Vibrotactile sensitivity and the frequency responseof the Pacinian corpuscle. Psychonomic Science 4, 1 (2014), 135–136. https://doi.org/10.3758/bf03342215